New Features and Enhancements

Facial Recognition modules

1. Integrated new STC Facial Recognition module 
2. Upgraded Tevian recognition module:
   1. Added an option to recognize emotion, facial attributes, wearing masks, fake faces, and introduced 3 new identification algorithms 
   2. 9 parameters that allow fine-tuning of the Tevian module are added to the registry 
   3. In online recognition, the best matching face is now displayed up to the moment when the module loses the track 
   4. Automatic adding of faces to DB now automatically ignores group photos.
   5. Optimized memory consumption by the Tevian recognition module 
3. Updated VisionLabs recognition module to version 3.6.3 
4. Removed the outdated VeriLook facial recognition module.
5. Integrated the ISD access and payment system 
6. KARS service is now supported in 64-bit mode 

Interface improvements

1. The Face Search tab now includes filters by name, department and similarity rate 
2. You can now choose between variants of facial search: within an exact period, or just "recently" 
3. Added an option to save specified facial search parameters (filters) 
4. You can now perform a fast search by double clicking the mouse button on a face displayed in the captured/recognized faces log. Search is performed with a specified filter 
5. Added filtering by department for searches in the Face Monitor archive 
6. Added an option of filtering by department when searching in the facial DB 
7. Added an option of filtering by camera when building a report on unique visitors 
8. Full names of recognized persons now appear over the facial bounding box 
9. Added an option to select a Video Surveillance Monitor where, after you left-click on a face in Monitor, Search or Analytics tab, an archive appears with the timeline positioned on a relevant fragment 
10. Facial monitoring UI now allows switching to a camera on the Map 
11. Added a warning message before deleting all faces 
12. Optimized facial search UI for operating on large facial databases 
13. You can now enter archives and search faces from the video analytics window 
14. Added minor improvements in displaying the table of recognized faces 
15. Revised logic of launching search: a search starts immediately from the context menu. If you load a facial image from a file, you have to click the Search button to start the search 
16. You can now use context menu to send a captured face to an external system for analysis 
17. You can now check a folder containing photos for faces that are present in a database 
18. Added simple mode in which facial recognition control panels and interface buttons are hidden 
19. Introduced an option to use spaces as delimiters when adding faces from files 
20. Facial search results interface now includes an option to increase column height along with facial images' dimensions. This setting is saved and automatically applied to subsequent searches 

New configuration capabilities

1. You can now choose between color or monochrome image to send it to the facial recognition module 
2. Added an option to specify recognition area in facial detector 
3. Added a Generate an Event on Recognition option which allows you to set the event source: Server or recognition channel 
4. During Face Intellect installation, port 10,000 is now added to firewall exclusions list 
5. In face search, the default similarity threshold is set to 80% 
6. Pixel resolution of video streams used in facial capture channels is now indicated in the debug window 
7. Added the Mark Similar Faces setting which allows grouping faces by similarity rate during capture for speeding up building reports on unique faces 
8. Renamed Online Analytics setting to Generating Access Event 
9. Added an Allow Department Creation checkbox that permits/forbids creation of a new department when adding a photo 
10. Implemented interoperability with the face synchronization module 

Face Intellect facial database

1. The default limit of recognized faces archive (for FIFO deletion) is reduced to 1000 MB 
2. Removed the Use.FaceDB = 0 mode (storing metadata, images, and FIRs in a database) 
3. For a Facial DB Replication object, the system now sequentially surveys all DBs responsible for system operation if one of the servers goes offline.
4. Added facial DB replication across servers to allow adding faces to all DBs by adding to just one database 
5. Optimized creating and loading facial DB indexes. Revamped index storage/writing/reading, optimized indexing.
6. You can now restore entry/exit data from saved indexes even if the FIR database is lost. To restore data, you can use the "Script database" command in the Debug window and the FaceRecognitionTool.exe utility 

Programming, configuration file, registry key and API functions

1. Added registry keys allowing setting the number of CPU cores used to generate FIRs 
2. The FIND_PERSON event now has parameters containing user ID upon creation and user ID in the Intellect system 
3. The PHOTO_ADDED command now has the "biggestFace" parameter allowing selection of the largest facial image when adding a group photo via a script 
4. The "face_client.run.config" file now contains the AutoSizeColumnHeaders setting which allows automatically adjusting column width to the UI window 
5. Added the Tevian.TrackBestFramesCount registry key which allows setting averaging of facial parameters by the best frame, or over a specified time period 
6. The FACE_LEAVE event now contains facial image quality parameter 
7. The FIND_PERSON event now contains age and gender parameters 
8. The RECOGNITION_PASS event now contains camera information 
9. The response to the GetAnalitycByTime API request now contains gender and emotions data 
10. Added the FaceTracker.LogTrackerQuality registry key allowing detailed logging of a facial detection tool 
11. The face_client.exe configuration file now contains the HideCheckBoxIntellectPerson setting which allows you to always set a unique Person_ID identifier when adding a face to the DB 
12. Added the GET_IMAGE reaction which allows returning an image in Base64 format 
13. The PERSON_NOT_FOUND event now contains the "confidence" parameter which characterizes the quality of recognition of a captured face 
14. Added capture timestamp parameters to the FIND_PERSON event 
15. Added a parameter containing similarity rate to the IDENTIFY_DECLINE event 

Other features

1. Face Intellect 7.3 is compatible with Intellect software versions starting from 4.11.0. No compatibility with older versions is provided.
2. Improved and upgraded logging in Face Intellect.
3. Introduced interoperability with Access Manager in the ACFA Intellect system.

Bug Fixes

1. Facial searches now start correctly from the context menu.
2. Fixed a bug that stopped facial capture on each 21st day.
3. Optimized and sped up Facial Recognition server operation.
4. Facial Recognition and Search UI now does not crash on selecting the Update Every 1 Second checkbox in the Analytics tab.
5. Facial Recognition and Search UI now does not crash on searching a face which was added from a file.
6. Corrected an error caused by damage to the index file during the start of re-generation of FIRs.
7. Fixed a bug that led to the distribution of CPU load across just 4 cores.
8. Fixed a bug with exporting faces whose names contain OS reserved characters.
9. Fixed an error that led to recognizing faces in non-facial objects.
10. Facial Recognition Server now does not crash on double verification.
11. Fixed a bug that led to selecting video streams unsuitable for face capture. The face capture stream now matches the video analytics stream setting.
12. Departments are now set correctly during mass adding of faces to the database.
13. Fixed an error that affected tracking of newly recognized faces.
14. Reduced latency in navigation across system objects after the installation of Face Intellect.
15. The Facial Recognition Server now does not crash, and memory leaks do not occur while using the VisionLabs recognition module.
16. The Facial Recognition Server now does not crash on exit if the Create FIR in Detection Unit function is set to Off.
17. The Facial Recognition Server now does not crash on starting facial recognition without a recognition module.
18. The Tevian Facial Recognition module now does not crash on AMD CPUs.
19. Fixed a bug that allowed adding more faces than the license permits.
20. The PATH system variable now is not damaged during the Face Intellect installation.
21. Fixed a rare software exception that occurred when adding faces to the DB.
22. Recognized faces are now displayed correctly in a distributed configuration.
23. Increased stability of Face Intellect operation on virtual machines.
24. Fixed a bug that added one extra entry/exit to analytics data.
25. Fixed an error that occurred when exporting large amount of data containing gender analytics results.
26. Face Intellect's facial database now synchronizes correctly with the Intellect users database.
27. Fixed a memory leak that occurred when adding faces to the DB.
28. Fixed a memory leak during face recognition.
29. Fixed a bug that included a face from the previous search in PDF reports.
30. Fixed an error that occurred when adding a face via the Video Footage.
31. The Facial Recognition Server now does not crash on recognizing a face of certain pixel dimensions.
32. Fixed a bug where a newly created department is not displayed in the Facial Recognition and Search UI.
33. Multiple face capture modules now work correctly on a single camera.
34. Facial Recognition module now does not crash on re-generation of FIRs via the Debug window if a facial image is damaged or removed from the database.
35. Fixed a bug that led to creation of new departments despite the "not allowed" setting upon mass adding of faces.